Generalized pollination system: Are floral traits adapted to different pollinators?

Andrea B. Lemaitre, Carlos F. Pinto & Hermann M. Niemeyer

Arthropod- Interactions An international journal devoted to studies on interactions of , mites, and other with

ISSN 1872-8855 Volume 8 Number 4

Arthropod-Plant Interactions (2014) 8:261-272 DOI 10.1007/s11829-014-9308-1

1 23 Your article is protected by copyright and all rights are held exclusively by Springer Science +Business Media Dordrecht. This e-offprint is for personal use only and shall not be self- archived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23 Author's personal copy

Arthropod-Plant Interactions (2014) 8:261–272 DOI 10.1007/s11829-014-9308-1

ORIGINAL PAPER

Generalized pollination system: Are floral traits adapted to different pollinators?

Andrea B. Lemaitre • Carlos F. Pinto • Hermann M. Niemeyer

Received: 29 January 2014 / Accepted: 20 May 2014 / Published online: 4 June 2014 Ó Springer Science+Business Media Dordrecht 2014

Abstract Flowers that are pollinated both during the day in -pollinated flowers. Flowers of E. chiloensis ssp. and at night could exhibit two different groups of chiloensis seem morphologically adapted to hawkmoth pollinators and produce two different sets of attractants and pollination, but diurnal and nocturnal pollinators contribute rewards. We explored the patterns of emission of flower to similar extents to reproductive success. Additionally, scents and production of nectar in the diurnal and nocturnal pollinators showed a synergic effect chiloensis ssp. chiloensis, in relation to the patterns of on the product of fruit set and seed set. The results are activity of its diurnal and nocturnal pollinators. We mea- discussed in terms of the linkage between floral traits and sured frequency of flower visitors, analyzed floral scents, perception abilities and requirements of pollinators. measured nectar production and sugar concentration, and performed pollination exclusion experiments. were Keywords Hawkmoth Bees Floral scents Nectar the main visitors at daytime and hawkmoths at nighttime. Pollination contribution Diurnal scents were dominated by several compounds that can attract a wide range of pollinators, whereas nocturnal scents were less diverse and were dominated by (E)-ner- Introduction olidol, a compound eliciting antennal responses in hawk- moths. Nectar volume and sugar concentration at night A pollination syndrome comprises the entire set of floral were similar to those recorded in hawkmoth-pollinated traits involved in the attraction of a particular pollinator flowers. Daytime nectar volume was higher than those (Proctor et al. 1996). Flowers with generalized pollination commonly found in bee-pollinated flowers, but similar to syndrome are those that attract more than one type of those found in flowers pollinated by several pollinators. pollinator (e.g., Young 2002; Schlumpberger and Badano Daytime sugar concentration was similar to those recorded 2005; Muchhala et al. 2008; Ortega-Baes et al. 2010; Schmid et al. 2010; Yokota and Yahara 2012) and are much more common than was thought until a few years ago Handling Editor: Lars Chittka. (Fleming et al. 2001). This syndrome may be an optimal strategy in highly variable environments, where the abun- Electronic supplementary material The online version of this article (doi:10.1007/s11829-014-9308-1) contains supplementary dance and distribution of pollinators fluctuate temporally material, which is available to authorized users. and spatially, because it prevents plant reproductive suc- cess to be negatively affected by the absence of particular & A. B. Lemaitre C. F. Pinto H. M. Niemeyer ( ) pollinators (Waser et al. 1996; Fleming et al. 2001). Departamento de Ciencias Ecolo´gicas, Facultad de Ciencias, Universidad de , Casilla 653, Santiago, Chile Different types of pollinators may have different per- e-mail: niemeyer@abulafia.ciencias.uchile.cl ception abilities and requirements (e.g., Dell’Olivo and Kuhlemeier 2013; Schemske and Bradshaw 1999; Briscoe C. F. Pinto and Chittka 2001; Chittka and Raine 2006; Kim et al. Laboratorio de Ecologı´a Quı´mica, Facultad de Ciencias y Tecnologı´a, Universidad Mayor de San Simo´n. Parque La Torre, 2011). For example, hawkmoths are mainly attracted by Cochabamba, olfactory cues (Yamamoto et al. 1969; Mechaber et al. 123 Author's personal copy

262 A. B. Lemaitre et al.

2002) and require large amounts of nectar with low sugar main visitors are bees. Because of the different perception concentration as reward (Opler 1983; Baker and Baker abilities and requirements of their floral pollinators, we 1983; Kaczorowski et al. 2005), and bees are mainly hypothesize that composition of flower scents, nectar vol- attracted by visual cues (Milet-Pinheiro et al. 2012) and umes, and sugar concentration in the nectar of E. chiloensis need lesser amounts of nectar with high sugar concentra- ssp. chiloensis differs between day and night, presenting tion (Opler 1983; Baker and Baker 1983). In this sense, characteristics related to the perception abilities and flowers that are pollinated both day and night may produce requirements of each group of pollinators. Furthermore, we flower scents (Dotterl et al. 2012) and nectar (Amorin et al. hypothesize that both groups of pollinators contribute to 2012) with different characteristics during the day and the reproductive success of the cactus. night, associated to their day and night pollinators, respectively. Moreover, for a generalized pollination syndrome to be Methods and materials considered as optimal, different groups of pollinators should contribute to the reproductive success of the plant Study area and study (Fleming et al. 2001); these contributions depend in general on both the visitation rate and the effectiveness in pollen This study was conducted from November 2012 to Feb- transfer (e.g., Thomson 2000; Aigner 2004). Different ruary 2013 at the Reserva Nacional Rı´o Clarillo (RNRC), pollinators may differ in their frequency of visits and located 45 km southeast of Santiago (33°410–33°510S, pollination effectiveness (Waser et al. 1996). Hawkmoths 70°240–70°290W) between 860 y 3057 m.a.s.l. The climate are generally considered very effective pollinators: Due to corresponds to a mediterranean regime with rainy winters their large body size and their foraging behavior, they are and dry summers (Niemeyer et al. 2002). During the study capable of carrying large amounts of pollen along great period, the mean day temperature was 19 ± 0.5 °C and distances (Johnson 1995; Johnson and Liltved 1997), and a mean relative humidity was 54 ± 3 % (data from the agro- single visit can successfully fertilize a given flower meteorological station at RNRC; mean of data taken at (Willmot and Bu´rquez 1996). In contrast, bees are con- 09:00, 14:00, and 18:00, during the 4 months that the study sidered ineffective pollinators since, unlike hawkmoths lasted). The vegetation of the area is dominated by ever- which are nectar collectors, most bees are pollen collectors green plants of the sclerophyllous forest and matorral and the pollen mass transported, being mixed with nectar vegetational region (Gajardo 1993). E. chiloensis ssp. and resin, is not all available for pollination (Williams and chiloensis is endemic to central Chile from La Serena Adam 2010). (29°020S) to Talca (36°330S) (Luebert and Pliscoff 2006) Several species of cacti from temperate regions exhibit and exhibits different pollination syndromes along its dis- generalized pollination syndromes; their main nocturnal tribution. In its northern distribution (Coquimbo: 31°300S, visitors are moths while their main diurnal visitors are bees 71°060W), it exhibits a diurnal pollination syndrome where (de Viana et al. 2001; Schlumpberger and Badano 2005; flowers open between 06:00 and 07:00 h, remain open an Schlumpberger et al. 2009; Ortega-Baes et al. 2010; average of 8.5 h and their main visitors are bees (Ossa and Alonso-Pedano and Ortega-Baes 2012). Since the identity Medel 2011). In contrast, in the southern range (Rancagua: of the main pollinators changes drastically within a short 34°100S, 70°430W), it exhibits a generalized pollination period of time, these flowers are ideal models to test syndrome where flowers open between 18:00 and 20:00 h, hypotheses of character adaptation to different pollinators remain open for 20–22 h on average and their main diurnal (e.g., Dotterl et al. 2012; Amorin et al. 2012). visitors are bees, and their main nocturnal visitors are In this paper, we explore the patterns of emission of moths (Walter 2010). These previous studies have deter- flower scents and nectar production in Echinopsis chilo- mined the reproductive system of this cactus as self- ensis ssp. chiloensis, in relation to the patterns of activity of incompatible and totally dependent on cross-pollination. its diurnal and nocturnal pollinators. Flowers of the Individuals of E. chiloensis ssp. chiloensis were chosen columnar cactus E. chiloensis ssp. chiloensis (Colla) H. from among the population growing along the main river at Friedrich and G. D. Rowley exhibit traits that conform to a the RNRC, within an area of approximately 25 ha. Only hawkmoth pollination syndrome (Dobson 2006), i.e., they one flower per individual was used. Anthesis in the study are large, cone shaped, white colored, without nectar area began mostly between 18:00 and 21:00 h, the mean guides, open at dusk and stay open for periods that vary time of opening being ca. 19:30; flowers remained open for from ca. 8 to 22 h depending on latitude. Indeed, Walter 20–22 h on average. During summer, dusk begins at (2010) described that the main nocturnal visitors and the around 21:00; in this scenario, potential diurnal pollinators major contributors to fruit set and seeds per fruit set in this have the chance to visit recently opened flowers up to 3 h cactus are hawkmoths, and also that during the day the before dusk. 123 Author's personal copy

Generalized pollination system 263

Floral visitors diameter 0.25 mm, and film thickness 0.25 lm). Helium was used as carrier gas with a flow rate of 1.3 ml/min. Three periods of observations during floral anthesis were Temperature of injection port and detector was set at defined: opening (18:00–21:00), night (21:00–06:00) and 250 °C. The oven temperature was programmed at 35 °C day (06:00–18:00). Thirteen different flowers that opened for 5 min and then increased by 5 °C every minute until around the same time in different days (between 18:30 and 200 °C was reached; this final temperature was maintained 20:00) were selected to conduct observations of floral for 10 min. Volatile compounds were identified according visitors. In daytime (opening and day periods), direct to their mass spectra and their retention indices using the observations were made for periods of 30 min each hour, NIST 2011 standard reference database and The Pherobase beginning with the opening of the flowers and ending with (El-Sayed 2012). their closing during the following day. The visiting insects were collected using an net, placed in Eppendorf Nectar production tubes and later identified by specialists (see Acknowledg- ments). In nighttime, the flowers were filmed with infrared In order to determine nectar production of E. chiloensis light for periods of 90 min every 3 h, and insects observed ssp. chiloensis along anthesis, nectar collection was per- during playback were identified by the same specialists. formed during the periods of highest visitor activity (see The mean visitation rate for each insect species was esti- ‘‘Results’’), within each period of anthesis; additionally, a mated as the number of individuals per hour per flower. collection was performed at dawn (06:30–08:30) to deter- The peak visiting activity hours during each of the obser- mine whether the nectar was produced only during the vation periods were determined to establish the times to night or also during the day. Fifty different flowers that collect nectar and floral scents so that these traits may be opened around the same time (18:30–20:00) in different related with visitors activity. cactus individuals were selected and individually bagged with tulle bags before opening to avoid nectar consump- Collection of floral scents tion. Two types of nectar measurements were performed. The first type involved 40 different flowers (ten flowers for Volatile scent collection was performed in fresh uncut each time period, i.e., opening, night, dawn and day), thus flowers in the field using the dynamic headspace adsorption obtaining the cumulative volume of nectar along anthesis method (Raguso and Pellmyr 1998). Six flowers that when flowers are not visited (accumulated nectar mea- opened around the same time (between 18:30 and 20:00) of surement). The second type of measurement was made in different cactus individuals were selected; the flowers were the remaining ten flowers repeatedly every 6 h, therefore, at all times individually enclosed in tulle bags, except when obtaining the nectar volume produced between hours they were submitted to scent collection. The collection (repeated nectar measurement). All nectar samples were system consisted of an inert oven bag (Melitta, Minden, collected and measured using 100 ll Hamilton syringes. Germany) which enclosed the flower, a silicone hose inlet, Sugar concentration in the nectar (% sucrose) was mea- a silicone hose outlet, and Teflon tape used for sealing the sured with a manual refractometer. For this, 40 ll of each system. Compressed synthetic air, manufactured by mixing nectar sample was used to cover completely the prism of pure oxygen (20 %) and nitrogen (80 %), was injected into the refractometer. When nectar volume of a single flower the inlet and extracted at the outlet to collect the volatile was less than 40 ll, a measured amount of distilled water compounds without getting contamination by compounds was added to a final volume of 50 ll; final concentrations in the natural air around the plant. Airflow at the inlet and were determined by applying the dilution factor to the outlet was adjusted to 250 ml/min, the first using the reg- measured concentration. ulator attached to a synthetic air tube and the second using an extraction pump (PAS-500), a manual flowmeter and Contribution of pollinators to reproductive success flow regulator. Floral scents were collected in Porapak Q columns attached to the bag outlet during the periods of One hundred flowers belonging to different cactus indi- highest visitor activity (see Results) within each period of viduals and opening at around the same time (between anthesis, each collection lasting 1 h. 18:30 and 20:00) were selected and assigned haphazardly In the laboratory, the Porapak Q columns were eluted to the following treatments: (1) total exclusion of pollina- with 1 ml of dichloromethane, and the extracts stored in tors: Flowers were excluded with tulle bags before opening 1.8-ml glass vials at -10 °C. The analysis of the volatiles until after closing (N = 25); (2) nocturnal exclusion of was performed on a gas chromatograph coupled to a mass pollinators: Flowers were bagged from 20:00 until 06:00 spectrometer (Shimadzu QP2010 Ultra). Samples were the next morning (N = 25); (3) diurnal exclusion of poll- analyzed in a capillary column (RTx-5MS: length 30 m, inators: Flowers were bagged from opening until 20:00 and 123 Author's personal copy

264 A. B. Lemaitre et al. from 06:00 until closing (N = 25); (4) without exclusion of Castellan 1988). A Spearman’s rank order correlation was pollinators: Flowers remained without bags (N = 25). The used to evaluate the relation between the volume produced flowers were monitored until abortion or fruit production; and sugar concentration of nectar measured using the the fruits produced were collected once they were mature accumulated method (Siegel and Castellan 1988). Analyses (after dehiscence) to determine the number of seeds pro- were performed in STATISTICA 8 (StatSoft Inc. 2006), duced per fruit. Subsequently, 396 seeds were selected and the graphics made using SIGMA PLOT 11.0 (Systat randomly per treatment (nocturnal exclusion, diurnal Software Inc. 2008). For easier interpretation of the results, exclusion, without exclusion, 1,188 in total), for use in the nectar untransformed values are shown. germination experiments under standard conditions: con- stant temperature of 23 °C and 14:10 light:darkness pho- toperiod for 30 days (e.g., Walter 2010). The seeds were Results separated in groups of 18 in Petri dishes with moist cotton and covered with aluminum paper. Floral visitors

Statistical analyses Twenty-four insect species belonging to five orders were observed visiting flowers (Table 1). Few visitors were Relative visitation rates and relative abundance of floral observed during the opening and night periods; visits were volatile compounds were calculated since raw values particularly scarce during the night period (Table 1). In showed substantial inter-individual variations. All data contrast, during the day, there were more visitor species were square root (datum ? 1) transformed because the data and visits were more frequent, being the matrix contained many zero values and to reduce the most frequently observed group. The most common species influence of particularly abundant species or volatile was the fly Dolichophaonia sp., followed by the bees Apis compounds (Quinn and Keough 2002). To compare the mellifera, Leioproctus semicyaneus, and Pseudagaposte- assembly of visitors and flower scent composition between mon pissisi (Table 1). The peak visiting activity occurred at time periods (opening, night and day), Bray and Curtis approximately the zeroth, sixth, and 18th hour of anthesis similarity indices were calculated and similarity analysis of and corresponded to 18:30–20:30, 00:30–02:30, and permutations (ANOSIM; Clarke and Gorley 2006) subse- 11:30–13:30 h, respectively. quently performed with 10,000 random permutations; the The assemblage of floral visitors varied significantly Bonferroni correction factor was applied to reduce type I between time periods (one-way ANOSIM: R = 0.37, error (Quinn and Keough 2002). To identify the species P \ 0.001). All pairwise comparisons were significantly and compounds responsible for the variability in visitation different between time periods (one-way ANOSIM: rate and scents between time periods, respectively, simi- opening vs. night, R = 0.06, P = 0.046; opening vs. day, larities of percentages (SIMPER) analyses were performed. R = 0.55, P \ 0.001; night vs. day, R = 0.54, P \ To show the variation of visitors and scents between time 0.001). According to the SIMPER analysis, the main periods, the nonmetric multidimensional scaling (NMDS) species responsible for differences between time periods method was applied based on the matrix of Bray–Curtis were: the fly Dolichophaonia sp., very frequent during the similarities (e.g., Balao et al. 2011). Analyses were per- opening and the day periods but absent at night; the bee formed in the statistical program PAST (Hammer et al. A. mellifera, present during the opening period, very often 2001). during the day period but absent at night; and the bees The number of fruits produced and the number of ger- L. semicyaneus, P. pissisi, and D. chilensis,presentwith minated seeds were compared between exclusion treat- high frequency only during the day period. ments through the chi-square test of association (Dytham 2011). Parametric one-way ANOVA was used to compare Floral scents nectar volume between periods and the number of seeds per fruit between exclusion treatments. Because these A total of 29 volatile compounds were identified in the variables did not follow a normal distribution, they were flower scents (Table 2), 21 (74.2 %) of which occurred transformed with a Box–Cox linear function [yk = (yk-1)/ during the three time periods (opening, night and day). ln yk, for (k = 0)] (Box and Cox 1964). A posteriori During the opening and night periods, scents were domi- pairwise comparisons were performed using Tukey’s tests nated by sesquiterpenes and monoterpenes; in contrast, (Dytham 2011). Nonparametric one-way Kruskal–Wallis during the day period, scents were dominated by aromatic test was used to compare nectar sugar concentrations esters (Table 2). between periods; this analysis was followed by nonpara- The composition of flower scents varied significantly metric Dunn test for unbalanced designs (Siegel and between time periods (one-way ANOSIM: R = 0.55, 123 Author's personal copy

Generalized pollination system 265

Table 1 Visitors, mean visitation rates (VR: visits/h), standard error (SE), and percentage of visits (%) to the cactus E. chiloensis ssp. chiloensis during three different time periods (opening, night, day) Order/species (family) Opening Night Day VR (SE) % VR (SE) % VR (SE) %

Coleoptera Bilyaxia concinna (Buprestidae) 0.31 (0.21) 0.59 Epiclines gayi (Cleridae) 0.15 (0.10) 0.30 Amecocerus sp. (Melyridae) 0.11 (0.11) 20.00 3.54 (1.25) 6.84 Arctodium vulpinu (Scarabaeidae) 0.46 (0.22) 0.89 Sub-totals 0.11 (0.11) 20.00 4.46 (1.57) 6.58 Diptera Dolichophaonia sp (Muscidae) 11.33 (1.45) 89.47 14.15 (5.36) 27.34 Mitrodetus dentitarsis (Mydidae) 0.23 (0.12) 0.45 Trichophthalma murina (Nemestrinidae) 0.15 (0.10) 0.30 Mycteronimyia conica (Tabanidae) 0.38 (0.18) 0.74 Seioptera importans (Ulidiidae) 0.38 (0.18) 0.74 Sub-totals 11.33 (1.45) 89.47 15.31 (5.3) 22.56 Hemiptera Leptoglossus chilensis (Coreidae) 0.67 (0.33) 5.26 0.22 (0.15) 40.00 0.15 (0.10) 0.30 Sub-totals 0.67 (0.33) 5.26 0.22 (0.15) 40.00 0.15 (0.10) 0.22 Hymenoptera Diadasia chilensis () 4.85 (1.56) 9.36 Apis mellifera (Apidae) 0.67 (0.67) 5.26 11.54 (4.36) 22.29 Manuelia gayi (Apidae) 3.31 (1.32) 6.39 Leioproctus semicyaneus (Colletidae) 10.08 (4.82) 19.47 Coenohalictus dolator (Halictidae) 1.69 (0.70) 3.27 Pseudagapostemon pissisi (Halictidae) 8.77 (3.40) 16.94 Lasioglossum herbstiellu (Halictidae) 1.85 (0.88) 3.57 Ruizantheda mutabilis (Halictidae) 1.62 (0.77) 3.12 Ruizantheda proxima (Halictidae) 0.85 (0.61) 1.63 Trichothurgus dubius (Megachilidae) 1.00 (0.53) 1.93 Megachile distingu¨enda (Megachilidae) 1.00 (0.34) 1.93 Unidentified Microhymenopteran 1.31 (0.66) 2.53 Sub-totals 0.67 (0.67) 5.26 47.85 (14.77) 70.64 Lepidoptera Manduca sexta (Sphingidae) 0.22 (0.18) 40.00 Sub-totals 0.22 (0.18) 40.00 Grand totals 12.67 (1.45) 100 0.56 (0.2) 100 48.92 (14.96) 100

P \ 0.001). The scents of the day period were significantly shows the clear differences in scent composition between different from those of the night period (one-way ANO- day and night and the similarity between opening and day SIM: R = 0.80, P = 0.007), while the differences between periods. the scents emitted during the opening period were mar- ginally significant with respect to the night period (one- Nectar production way ANOSIM: R = 0.46, P = 0.044) and to the day per- iod (R = 0.34, P = 0.052). According to SIMPER Nectar was available at the opening of flowers and was analysis, the main compounds responsible for differences produced continuously along anthesis (Fig. 2A). The between time periods were the terpenes (E)-nerolidol, (E)- accumulated nectar volume varied significantly along floral ocimene and b-pinene and the aromatic ester methyl ben- anthesis (ANOVA: F = 17.370, P \ 0.001): Nectar accu- zoate. The graph resulting from the NMDS method (Fig. 1) mulated during the whole anthesis period (collected during

123 Author's personal copy

266 A. B. Lemaitre et al.

Table 2 Frequency (N), average relative proportions (area %), and standard error (SE) of 29 compounds present in the floral scents of six flowers of ssp. chiloensis during each of three different time periods (opening, night, day) Compound class/name Retention index Opening Night Day N Area % (SE) N Area % (SE) N Area % (SE)

(a) Terpenes Monoterpenes a-Pinene 931 6 1.10 (2.25) 4 0.28 (0.57) 6 0.84 (0.88) Camphene 947 3 6.04 (13.78) 3 1.17 (3.16) 2 2.03 (0.96) b-Pinene 975 6 1.51(2.35) 5 0.32 (0.62) 6 3.01 (6.66) b-Myrcene 990 1 0.01(1.00) 3 0.42 (0.13) 3-Carene 1,006 5 0.30 (0.91) 2 0.08 (0.29) 4 1.10 (2.57) m-Cymene 1,024 5 0.15 (0.96) 2 0.05 (0.16) 4 1.03 (1.37) Limonene 1,027 6 0.94 (2.09) 4 0.47 (1.15) 5 3.74 (0.85) (Z)-b-Ocimene 1,039 5 1.06 (4.05) 5 4.31 (2.17) 5 0.93 (2.43) (E)-Ocimene 1,050 5 18.49 (11.60) 5 38.73 (11.17) 6 5.65 (7.75) Linalool 1,099 1 0.01 (0.00) 4 2.51 (1.09) Neo-allo-ocimene 1,131 1 0.10 (0.00) 2 0.69 (0.38) 2 0.79 (4.46) Sub-totals 29.71 (3.21) 49.04 (6.47) 19.13 (0.92) Sesquiterpenes 2,6,11-Trimethyldodecane 1,281 3 0.17 (0.91) 3 0.04 (0.05) 3 0.72 (0.12) a-Cubebene 1,346 1 0.05 (0.00) b-Farnesene 1,462 2 1.52 (0.61) 5 0.26 (0.10) 1 1.08 (1.00) Germacrene D 1,489 2 0.28 (6.50) 2 0.94 (3.04) 2 0.06 (0.18) a-Farnesene 1,513 4 15.70 (3.28) 6 4.85 (2.12) 3 7.57 (2.35) d-Cadinene 1,531 1 0.13 (1.00) 2 0.23 (0.50) 1 0.01 (1.00) (E)-Nerolidol 1,570 2 36.13 (9.15) 6 42.05 (6.85) 5 9.85 (2.00) Sub-totals 53.97 (5.21) 48.36 (6.33) 19.29 (1.64) (b) Benzenoids Aliphatic benzenoid p-Xylene 871 1 0.005 (1.00) 2 2.46 (3.60) 4 0.21 (0.04) Sub-totals 0.005 (1.00) 2.46 (3.60) 0.21 (0.04) Aromatic esters Methyl benzoate 1,098 1 15.00 (0.00) 6 47.21 (12.57) Methyl salicylate 1,198 1 0.43 (0.00) 1 8.15 (0.00) Benzyl benzoate 1,769 1 1.66 (0.00) Sub-totals 15.43 (5.95) 57.02 (14.23) (c) Alkanes 3,4-Dimethylheptane 865 2 0.04 (0.45) 2 0.04 (0.17) Nonane 896 4 0.38 (0.58) 1 0.10 (0.00) 2 0.14 (0.04) Undecane 1,097 1 0.08 (0.00) Dodecane 1,199 1 0.03 (0.00) Tetradecane 1,400 1 0.005 (0.00) 1 0.04 (0.00) 1 0.69 (0.00) Hexadecane 1,599 1 2.23 (0.00) Sub-totals 0.42 (0.08) 0.14 (0.02) 3.21 (0.35) (d) Aldehyde Decanal 1,206 1 0.46 (0.00) 1 1.15 (1.00) Sub-totals 0.46 (0.00) 1.15 (1.00) Grand totals 100 100 100

123 Author's personal copy

Generalized pollination system 267

A

B

Fig. 1 Non-metric multidimensional scaling (NMDS) based on Bray–Curtis similarity matrices of floral scents, in flowers of the cactus E. chiloensis ssp. chiloensis at three times during anthesis (2D stress value 0.298). Each point in space is related to a singular pattern of scents of a flower during a particular collection event (maximum N = 6) the day) was significantly higher than that accumulated from opening to night and at opening (Tukey: day vs. night, q = 5.511, P = 0.002; day vs. opening, q = 9.502, P \ 0.001). On the other hand, nectar accumulated from Fig. 2 A Nectar volume (ll, mean ± 1 SE) and B sugar concentra- tion (% sucrose, mean ± 1 SE) in flowers of the cactus E. chiloensis opening to dawn and from opening to night was signifi- ssp. chiloensis measured at three different periods within anthesis cantly higher than that accumulated at opening (Tukey: using two different methodologies (open circle accumulated nectar dawn vs. opening, q = 7.460, P \ 0.001; night vs. open- measurements; filled circle repeated nectar measurements). Different ing, q = 3.992, P = 0.037). Nectar production determined letters (a, b, c and A, B, C) indicate significant differences between periods at P = 0.05 through the repeated measurements method varied signifi- cantly along floral anthesis (ANOVA: F = 8.336, P \ 0.001): A higher amount of nectar was produced Flowers without exclusion produced a significantly higher during the day than during the night and opening periods proportion of fruit compared to flowers with nocturnal and (Tukey: day vs. night, q = 5.412, P = 0.003; day vs. diurnal exclusions (chi square: v2 = 5.56, P = 0.038 and opening, q = 6.024, P \ 0.001) and during dawn the v2 = 18, P \ 0.001, respectively). The proportion of fruit amount of nectar produced was higher than that produced produced by flowers with nocturnal exclusion was signifi- before opening (Tukey: q = 4.219, P \ 0.025). cantly higher than by flowers with diurnal exclusion (chi Sugar concentration was significantly different among square: v2 = 4.35, P = 0.037). Number of seeds set per periods (Kruskal–Wallis: H = 12.753, P = 0.005), being fruit varied significantly between floral exclusions higher at opening compared with the day period (Dunn’s: (ANOVA: F = 6.825, P = 0.003): Flowers without Q = 3.361, P \ 0.05). Other pairwise comparisons were exclusion produced a significant higher number of seeds not significantly different (Fig. 2B). Sugar concentration per fruit in comparison with nocturnal and diurnal exclu- was only determined on samples obtained from the nectar sions (Tukey: q = 3.407, P = 0.045 and q = 4.737, accumulated measures experiment and was negatively P = 0.006, respectively) (Table 3); however, there was no correlated with nectar volume (Spearman: rs =-0.511, difference between diurnal and nocturnal exclusion treat- P = 0.001). ments (Tukey: q = 2.106, P = 0.309). The number of germinated seeds from flowers without exclusion was Contribution of pollinators to reproductive success significantly higher than those from the nocturnal and diurnal exclusions (chi square: v2 = 11.07, P \ 0.001 and Flowers that were excluded since their opening until their v2 = 9.96, P = 0.001, respectively); however, there was closing (total exclusion) did not produce fruit (Table 3). no difference in the number of germinated seeds between 123 Author's personal copy

268 A. B. Lemaitre et al.

Table 3 Fruit set, seed set (per fruit), Fruit set 9 seed set, and germinated seeds from flowers of the cactus E. chiloensis ssp. chiloensis under different pollinator exclusion experiments Exclusion treatment Fruit set Seed set Fruit set 9 seed set Germinated seeds (Proportion) N° (SE) N° N° ± SE (%)

Total exclusion 0 0 0 0 Without exclusion 0.80 a 2,850.2 (393.8) a 2,280 17.3 ± 0.2 (96.0) a Nocturnal exclusion 0.48 b 764.8 (221.1) b 367 16.2 ± 0.3 (89.9) b Diurnal exclusion 0.20 c 1,599.0 (376.8) b 320 16.3 ± 0.4 (90.7) b SE standard errors Different letters within columns indicate significant differences at P = 0.05 diurnal and nocturnal exclusions (chi square: v2 = 0.13, significant. In contrast, Amecocereus sp. could be acting as P = 0.72) (Table 3). The product of fruit set 9 seed set a pollinator because many species of this feed on was similar for the diurnal and nocturnal exclusion treat- pollen and can transport it on their pilose bodies (Smith- ments, and they were smaller than that for the no exclusion Ramı´rez and Ya´n˜ez 2010). However, the main nocturnal treatment (Table 3). pollinator of E. chiloensis ssp. chiloensis is likely to be M. sexta because hawkmoths are considered in general to be very effective pollinators due to their large body size, their Discussion foraging behavior, and their capability of carrying large amounts of pollen along great distances (Johnson 1995; During flower opening and the daytime period, the fly Johnson and Liltved 1997); additionally, M. sexta has been Dolichophaonia sp. was the most frequent visitor; other described as a frequent visitor of cacti (Raguso et al. 2003; frequent visitors during the daytime were the bees A. Schlumpberger et al. 2009; Ortega-Baes et al. 2010) and mellifera, L. semicyaneus, and P. pissisi. Dolichophaonia was observed previously by chance inside one of the sp. did not make contact with the flower anthers, and no flowers of E. chiloensis ssp. chiloensis (Walter 2010). pollen was observed on its body. Moreover, some larvae of Although compositional variation of flower scents along Dolichophaonia sp. were observed inside withered flowers; anthesis could simply be a consequence of circadian therefore, this fly is likely to visit flowers mainly in search of changes in light and temperature (Kolosova et al. 2001), places to lay eggs and, if it has an effect on pollination, it the present variations were consistent with the differential might not be significant. On the other hand, all bee species requirements of daytime as well as nighttime pollinators, as observed belong to pollen feeding families (Grimaldi and has been reported in other systems (Kaczorowski et al. Engel 2005; Williams and Adam 2010). Moreover, A. 2005, Dotterl et al. 2012). Diurnal scents were composed mellifera has been described as a frequent visitor to several by a large number of compounds; this is frequently the cactus species (Bustamante et al. 2010; Ortega-Baes et al. situation in flowers pollinated by different types of 2010; Larrea-Alca´zar and Lo´pez 2011; Alonso-Pedano and pollinators. Scents produced by bee-pollinated flowers are Ortega-Baes 2012), some species of the genus Diadasia highly variable and may be dominated by benzenoids, have been described as pollen collector specialists from cacti terpenes, or fatty acid derivatives (Dobson 2006). Diurnal (Sipes and Tepedino 2005), and Manuelia gayi has been scents of E. chiloensis ssp. chiloensis were dominated by reported as a pollinator of several plant species in the area the aromatic esters methyl benzoate and methyl salicylate, where the present studies were performed (Smith-Ramı´rez and the terpenes (E)-nerolidol, a-farnesene and (E)-ocim- and Ya´n˜ez 2010). Therefore, bees are likely to be the main ene. All these compounds were found in the cactus daytime pollinators of E. chiloensis ssp. chiloensis. Pereskia aculeate, which is mainly pollinated by bees and The only visitor species observed during the night were butterflies (Kaiser and Tollsten 1995). Moreover, (E)-ner- the hemipteran Leptoglossus chilensis, the beetle Amec- olidol, a-farnesene, and (E)-ocimene are common com- ocerus sp., and the hawkmoth Manduca sexta. The hemi- pounds in bee and moth-pollinated flowers (Kaiser and pteran is a generalist phytophagous insect (Levin 2000) Tollsten 1995; Raguso et al. 2003; Jurgens 2004; Grajales- that has not been reported as pollinator and that was not Conesa et al. 2011), and methyl benzoate and methyl observed making contact with the flower anthers, most of salicylate are commonly found both day and night in bee, the observations in the present study corresponding to moth, bat, and butterfly-pollinated flowers (Knudsen and mating events; hence, this insect most likely visits the Tollsten 1993; Kaiser and Tollsten 1995; Kolosova et al. cactus flowers as a herbivore or in search of a mate and 2001; Pettersson et al. 2004; Dobson 2006; Schlumpberger therefore, if it has an effect on pollination, it might not be and Raguso 2008; Grajales-Conesa et al. 2011).

123 Author's personal copy

Generalized pollination system 269

The less variable and less complex nocturnal scents of cacti (Fleming et al. 1996; de Viana et al. 2001; E. chiloensis ssp. chiloensis suggest they are possibly Schlumpberger and Badano 2005; Ortega-Baes et al. 2010). attracting a narrower range of pollinators. Scents of These results suggest that although the nighttime nectar hawkmoth-pollinated flowers are generally dominated by seems to be related to moths’ requirements, the increased terpenes, aromatic esters, and/or nitrogen compounds volume of nectar during the day seems to be of advantage (Kaiser and Tollsten 1995; Raguso and Willis 2003; to a large number of pollinators. Dobson 2006). Nocturnal scents of E. chiloensis ssp. Similarly to other temperate cactus species (Fleming chiloensis were almost completely dominated by terpenes, et al. 2001; Dar et al. 2006; Ortega-Baes et al. 2010; the most abundant being the sesquiterpene (E)-nerolidol Alonso-Pedano and Ortega-Baes 2012), in the pollination and the monoterpene (E)-ocimene, both commonly, but not system of E. chiloensis ssp. chiloensis, its different groups exclusively, found in flowers pollinated by hawkmoths of pollinators showed an approximately additive effect on (Knudsen and Tollsten 1993; Jurgens et al. 2003; Raguso fruit set and also on seed set per fruit. On the other hand, et al. 2003; Dobson 2006; Schlumpberger and Raguso the product of fruit set and seed set gave similar results 2008). Although we did not find aromatic esters or nitrogen under diurnal and nocturnal pollination, suggesting that compounds in nocturnal scents of E. chiloensis ssp. both groups of pollinators contribute to similar extents to chiloensis, the presence of (E)-nerolidol, an elicitor of the reproductive success of the cactus, presumably because antennal responses in several hawkmoth species (Raguso nocturnal pollinators are very effective and diurnal et al. 1996; Raguso and Light 1998) including M. sexta pollinators very frequent. This type of compensation in (Fraser et al. 2003), is of determining importance. pollinator activity affecting plant fitness has been reported Although volume and concentration variation of nectar in different contexts, e.g., diurnal and nocturnal pollinators along anthesis could simply be a consequence of physical affecting the product of fruit set 9 seed set (Wolff et al. conditions of the plant and environmental conditions 2003), a bumblebee and a hummingbird affecting also the (Willmer 2011), the present variations were consistent with product of fruit set x seed set (Waser and Price 1990), and a the differential requirements of daytime and nighttime bumblebee and a hummingbird affecting plant fitness pollinators, as has been reported in other systems (Morse through color choice (Waser and Price 1981). Moreover, and Fritz 1983; Macior 1986; Amorin et al. 2012). Diurnal the joint effect of diurnal and nocturnal pollinators on the nectar volumes in E. chiloensis ssp. chiloensis flowers were product of fruit set and seed was over three times higher higher than those observed in other bee-pollinated flowers that the added effects of the groups measured individually (ca. 115 ll observed values; vs. 10 ll, Opler 1983;15ll, (Table 3), suggesting a synergistic effect between the two Perret et al. 2001;54ll, Ossa and Medel 2011), but closer groups of pollinators. Reports in the literature have shown to the range recorded in other cactus flowers with gen- that interspecific interactions can alter the behavior of eralized pollination syndrome (ca. 115 ll, observed values; pollinators leading to increased pollination efficiency 110 ll, Silva and Sazima 1995; 129 ll, de Viana et al. (Greenleaf and Kremen 2006; Brittain et al. 2013). 2001; 170 ll; Schlumpberger et al. 2009). Sugar concen- Flowers of E chiloensis ssp. chiloensis show external tration in E. chiloensis ssp. chiloensis during the day was traits normally associated to hawkmoth pollination; hence, within the wide range reported for other bee-pollinated a higher contribution to reproductive success of nocturnal species (ca. 28 % sucrose w/w, observed values; pollinators over diurnal pollinators may be expected 21.2–77 %, Waser and Price 1981; 29 %, Perret et al. (Sletvold et al. 2012). Table 3 shows that this was not the 2001). On the other hand, nectar volume and sugar con- case. A possible explanation for this apparent contradiction centration in flowers of E. chiloensis ssp. chiloensis at the is that flowers may be morphological adapted to their night period suggest nectar as a reward trait related to ‘‘most effective pollinators’’ in past ecological contexts moths’ requirements. Values observed in this study were which seem to different to those presently prevailing (Ol- similar to those found for other cactus species pollinated by lerton 1996). In such cases, current main pollinators may nocturnal moths, both in nectar volumes (ca. 26 ll, not have affected the genetic integration of floral adapta- observed values; 23–29 ll, Ortega-Baes et al. 2010) as well tions that evolved in the past (Stebbins 1970). The latter as in sugar concentration (ca. 30 % w/w, observed values; could occur if current main pollinators are attracted and are 8.2–43.9 %, Pyke and Waser 1981; 15–30 %, Baker and effective over a wide range of forms of floral traits Baker 1983; Kaczorowski et al. 2005; 24–28 %, Fleming including the narrower range of forms fitted to the former et al. 1996; 25–29 %, Raguso et al. 2003; 20–30 %, Torres main pollinators (Aigner 2001). Apparently, current main et al. 2013). In E. chiloensis ssp. chiloensis, nectar is pollinators of E. chiloensis ssp. chiloensis are hawkmoths produced continuously both during day and night, the and bees, and bees that seem to be attracted by a wide greatest production and volume being observed during the range of morphological floral traits, included among which day and not at night as are common in moth-pollinated are the particular morphological floral traits that attract 123 Author's personal copy

270 A. B. Lemaitre et al. hawkmoths. The floral tube of E. chiloensis ssp. chiloensis indebted to the Organization of American States for a partial fel- flowers measured 13.25 ± 0.2 cm while the length of the lowship. We thank the generous and constructive comments and suggestions made by the reviewers of the original manuscript. proboscis of M. sexta ranges from 6 to 10 cm (Schlump- berger et al. 2009; Walter 2010); this forces the hawkmoth to penetrate into the flower to reach the nectar and in this References way remove and deposit large amounts of pollen (Gibson 2001). On the other hand, the floral tube is sufficiently wide Aigner PA (2001) Optimality modeling and fitness trade-offs: when to allow bees to enter and reach the nectar. Exerted stamens should plants become pollinator specialists? Oikos 95:177–184 of E. chiloensis ssp. chiloensis flowers also facilitate pollen Aigner PA (2004) Floral specialization without trade-offs: optimal removal by hawkmoths, which hover just at the stamens corolla flare in contrasting environments. Ecology 85:2560–2569 level when feeding on nectar (Grant and Grant 1983). This Alonso-Pedano M, Ortega-Baes P (2012) Generalized and comple- mentary pollination system in the Andean cactus Echinopsis type of stamens can also attract bees collecting pollen schickendantzii. Plant Syst Evol 298:1671–1677 (Miyake et al. 1998). Amorin F, Galetto L, Sazima M (2012) Beyond the pollination While morphological floral traits tend to be rather constant syndrome: nectar ecology and the role of diurnal and nocturnal within a species (Worley and Barrett 2000;Herrera2001), pollinators in the reproductive success of Inga sessilis (Faba- ceae). Plant Biol 15:317–327 nectar production and composition and floral scents are Baker HG, Baker I (1983) Floral nectar sugar constituents in relation physiological traits which depend on the physical conditions to pollinator type. In: Jones CE, Little RJ (eds) Handbook of of the plant and environmental conditions, which may change experimental pollination biology. Scientific and Academic over short spatial and temporal scales (Rathcke 1992;Torres Editions, New York, pp 117–141 Balao F, Herrera J, Talavera S, Do¨tterl S (2011) Spatial and temporal and Galetto 1998; Lake and Hughes 1999; Biernaskie and patterns of floral scent emission in Dianthus inoxianus and Cartar 2004; Longo and Fischer 2006; Macukanovic-Jocic electroantennographic responses of its hawkmoth pollinator. et al. 2004; Valtuena et al. 2007; Ono et al. 2008). In this way, Phytochemistry 72:601–609 flowers of E. chiloensis ssp. chiloensis may produce nectar Biernaskie JM, Cartar RV (2004) Variation in rate of nectar production depends on floral display size: a pollinator manip- and floral scents with characteristics which match the ulation hypothesis. Funct Ecol 18:125–129 requirements and perception abilities of diurnal and nocturnal Box GEP, Cox DR (1964) An analysis of transformation. J R Stat Soc pollinators without interference from each other. Ser B 26:211–243 In conclusion, flowers of E. chiloensis ssp. chiloensis Briscoe AD, Chittka L (2001) The evolution of color vision in insects. Annu Rev Entomol 46:471–510 show external traits normally associated to hawkmoth Brittain C, Williams N, Kremen C, Klein AM (2013) Synergistic pollination but also allowing bee pollination; moreover, the effects of non-Apis bees and honey bees for pollination services. floral scents and nectar rewards of E. chiloensis ssp. Proc Natl Acad Sci 280:19653–19656 chiloensis during the day and at night have characteristics Bustamante E, Casas A, Burquez A (2010) Geographic variation in reproductive success of Stenocereus thurberi (Cactaceae): consistent with the requirements and perception abilities of effects of pollination timing and pollinator guild. Am J Bot diurnal (bees) and nocturnal pollinators (hawkmoths), 97:2020–2030 respectively. Furthermore, the pollination system of E. Chittka L, Raine NE (2006) Recognition of flowers by pollinators. chiloensis ssp. chiloensis involves diurnal and nocturnal Curr Opin Plant Biol 9:428–435 Clarke KR, Gorley RN (2006) Primer v6: user manual/tutorial. pollinators contributing to similar extents to its reproduc- Primer-E Ltd, Plymouth tive success. These results contrast with those of Walter Dar S, Coro M, Valiente-Banuet A (2006) Diurnal and nocturnal (2010), who found that nocturnal pollinators contributed pollination of Marginatocereus marginatus (Pachycereeae: more than diurnal ones to the reproductive success of a Cactaceae) in Central Mexico. Ann Bot 97:423–427 de Viana ML, Ortega Baes P, Saravia M, Badano EI, Schlumpberger population of E. chiloensis ssp. chiloensis, and those of B (2001) Biologı´a floral y polinizadores de Trichocereus Ossa and Medel (2011), who found populations of E. pasacana (Cactaceae) en el Parque Nacional Los Cardones, chiloensis ssp. chiloensis with strictly diurnal pollination . Rev Biol Trop 49:279–285 syndrome. Additional studies exploring the variation in Dell’Olivo A, Kuhlemeier C (2013) Asymmetric effects of loss and gain of a floral trait on pollinator preference. Evolution 67:3023–3031 abundance and distribution of pollinators, their relative Dobson H (2006) Relationship between floral fragrance composition effectiveness and efficiency as pollinators, their require- and type of pollinator. In: Dudareva N, Pichersky E (eds) ments and perception abilities, and the relation of these Biology of floral scent. CRC Press, Boca Raton, pp 147–198 factors with floral traits are needed to understand the Dotterl S, Jahreib K, Jhumur US, Jurgens A (2012) Temporal variation of flower scent in Silene otites (Caryophyllaceae): a geographical variation of this cactus–pollinator interaction. species with a mixed pollination system. Bot J Linn Soc 169:447–460 Acknowledgments We thank Prof. Elizabeth Chiappa for identifi- Dytham C (2011) Choosing and using statistics: a biologist’s guide. cation of Hymenoptera, Andre´s Fierro for identification of Coleop- SPi Publisher Services, India tera, and Prof. Christian Gonza´les for identification of Diptera. The El-Sayed AM (2012) The pherobase: database of pheromones and support of LANBIO (Latin American Network for Research in Bio- semiochemicals. http://www.pherobase.com. Accessed 4 April active Natural Compounds) is gratefully acknowledged. ABL is 2013 123 Author's personal copy

Generalized pollination system 271

Fleming TH, Tuttle MD, Horner MA (1996) Pollination biology and Longo JM, Fischer E (2006) Effect of nectar secretion rate on the relative importance of nocturnal and diurnal pollinators in pollination and seed production by Passiflora speciosa Gardn. three species of Sonoran desert columnar cacti. Southwest Nat flowers (Passifloraceae) in the Pantanal. Rev Bras Bot 41:257–269 29:481–488 Fleming TH, Shaley CT, Holland JN, Nason JD, Hamrick JL (2001) Luebert F, Pliscoff P (2006) Sinopsis Bioclima´tica y Vegetacional de Sonoran desert columnar cacti and the evolution of generalized Chile. Editorial Universitaria, Santiago de Chile pollination systems. Ecol Monogr 71:511–530 Macior LW (1986) Floral resource sharing by bumblebees and Fraser AM, Mechaber WL, Hildebrand JG (2003) Electroantenno- hummingbirds in Pedicularis (Scrophulariaceae) pollination. graphic and behavioral responses of the sphinx moth Manduca Bull Torrey Bot Club 113:101–109 sexta to host plant headspace volatiles. J Chem Ecol Macukanovic-Jocic M, Duletic-Lausevic S, Jocic G (2004) Nectar 29:1813–1833 production in three melliferous species of Lamiaceae in natural Gajardo R (1993) La vegetacio´n natural de Chile: clasificacio´ny and experimental conditions. Acta Vet 54:475–487 distribucio´n geogra´fica. Universitaria, Santiago Mechaber WL, Capaldo CT, Hildebrand JG (2002) Behavioural Gibson AC (2001) The plants that love hawkmoths. Newsletters of responses of adult female tobacco hornworms, Manduca sexta, the Mildred E. Mathias Botanical Garden 4. http://www.botgard. to hostplant volatiles change with age and mating status. J Insect ucla.edu/Theplantsthatlovehawkmoths. Accessed 15 April 2014 Sci 2:5 (insectscience.org/2.5) Grajales-Conesa J, Mele´ndez-Ramı´rez V, Cruz-Lo´pez L (2011) Milet-Pinheiro P, Ayasse M, Schlindwein C, Dobson HEM, Dottler S Aromas florales y su interaccio´n con los insectos polinizadores. (2012) Host location by visual and olfactory floral cues in an Rev Mex Biodivers 82:1356–1367 oligolectic bee: innate and learned behavior. Behav Ecol Grant V, Grant KA (1983) Behavior of hawkmoths on flowers of 23:531–538 Datura meteloides. Bot Gaz 144:280–284 Miyake T, Yamaoka R, Yahara T (1998) Floral scents of hawkmoth- Greenleaf SS, Kremen C (2006) Wild bees enhance honey bees’ pollinated flowers in Japan. J Plant Res 111:199–205 pollination of hybrid sunflower. Proc Natl Acad Sci Morse DH, Fritz RS (1983) Contributions of diurnal and nocturnal 103:13890–13895 insects to the pollination of common milkweed (Asclepias Grimaldi D, Engel MS (2005) Evolution of the insects. Cambridge syriaca L.) in a pollen-limited system. Oecologia 60:190–197 University Press, Cambridge Muchhala N, Caiza A, Vizuete JC, Thomson JD (2008) A generalized Hammer Q, Harper DAT, Ryan PD (2001) PAST: paleontological pollination system in the tropics: bats, birds and Aphelandra statistics software package for education and data analysis. acanthus. Ann Bot 103:1481–1487 Palaeontol Electron 4:9 Niemeyer HM, Bustamante RO, Simonetti JA, Teillier S, Fuentes- Herrera J (2001) The variability of organs differentially involved in Contreras E, Mella JE (2002) Historia Natural de la Reserva pollination, and correlation of traits in Genisteae (Leguminosae: Nacional Rı´o Clarillo: un Espacio Para Aprender Ecologı´a. Papilionoideae). Ann Bot 88:1027–1037 Impresos Socias, Santiago Johnson SD (1995) Observations of hawkmoth pollination in the NIST-The National Institute of Standards and Technology (2011) South African orchid Disa cooperi. Nord J Bot 15:121–125 U.S. Secretary of Commerce on behalf of the United States of Johnson SD, Liltved WR (1997) Hawkmoth pollination of Bonatea America speciosa (Orchidaceae) in a South African coastal forest. Nord J Ollerton J (1996) Reconciling ecological processes with phylogenetic Bot 17:5–10 patterns: the apparent paradox of plant–pollinator systems. Jurgens A (2004) Flower scent composition in diurnal Silene species J Ecol 84:767–769 (Caryophyllaceae): phylogenetic constraints or adaption to Ono A, Dohzono I, Sugawara T (2008) Bumblebee pollination and flower visitors? Biochem Syst Ecol 32:841–859 reproductive biology of Rhododendron semibarbatum (Erica- Jurgens A, Witt T, Gottsberger G (2003) Flower scent composition in ceae). J Plant Res 121:319–327 Dianthus and Saponaria species (Caryophyllaceae) and its Opler PA (1983) Nectar production in a tropical ecosystem. In: relevance for pollination biology and . Biochem Syst Bentley B, Elias T (eds) The biology of nectaries. Columbia Ecol 31:345–357 University Press, New York, pp 30–79 Kaczorowski RL, Gardener MC, Holtsford TP (2005) Nectar traits in Ortega-Baes P, Saravia M, Su¨hring S, Godı´nez-Alvarez H, Zamar M Nicotiana section Alatae (Solanaceae) in relation to floral traits, (2010) Reproductive biology of (Cacta- pollinators and mating system. Am J Bot 92:1270–1283 ceae): the role of nocturnal and diurnal pollinators. Plant Biol Kaiser R, Tollsten L (1995) An introduction to the scent of cacti. 13:33–40 Flavour Fragr J 10:153–164 Ossa CG, Medel R (2011) Notes on the floral biology and pollination Kim W, Gilet T, Bush JWM (2011) Optimal concentrations in nectar syndrome of Echinopsis chiloensis (Cactaceae) in a population feeding. Proc Natl Acad Sci 4:16618–16621 of semiarid Chile. Gayana Bot 68:213–218 Knudsen JT, Tollsten L (1993) Trends in floral scent chemistry in Perret M, Chautems A, Spichiger R, Peixoto M, Savolainen V (2001) pollination syndromes: floral scent composition in moth-polli- Nectar sugar composition in relation to pollination syndromes in nated taxa. Bot J Linn Soc 113:263–284 Sinningieae (Gesneriaceae). Ann Bot 87:267–273 Kolosova N, Gorenstein N, Kish CM, Dudareva N (2001) Regulation Pettersson S, Ervik F, Knudsen JT (2004) Floral scent of bat- of circadian methyl benzoate emission in diurnally and noctur- pollinated species: West Africa vs. the New World. Biol J Linn nally emitting plants. Plant Cell 13:2333–2347 Soc 82:161–168 Lake JC, Hughes L (1999) Nectar production and floral characteristics Proctor M, Yeo P, Lack A (1996) The natural history of pollination. of Tropaeolum majus L. grown in ambient and elevated carbon Harper Collins Publishers, New York dioxide. Ann Bot 84:535–541 Pyke GH, Waser NM (1981) The production of diluted nectar by Larrea-Alca´zar DM, Lo´pez RP (2011) Pollination biology of hummingbird and honeyeater flowers. Biotropica 13:260–270 Oreocereus celsianus (Cactaceae), a columnar cactus inhabiting Quinn G, Keough M (2002) Experimental design and data analysis for the high subtropical . Plant Syst Evol 295:129–137 biologists. Cambridge University Press, Cambridge Levin P (2000) Leaf-footed bugs (Coreidae). In: Schaefer CW, Raguso RA, Light DM (1998) Electroantennogram responses of male Panizzi AR (eds) Heteroptera of economic importance. CRC Sphinx perelegans hawkmoths to floral and ‘greenleaf volatiles’. Press, Boca Raton, pp 337–403 Entomol Exp Appl 86:287–293

123 Author's personal copy

272 A. B. Lemaitre et al.

Raguso RA, Pellmyr O (1998) Dynamic headspace analysis of floral Systat Software Inc. (2008) SigmaPlot statistics user’s guide, version volatiles: a comparison of methods. Oikos 81:238–254 11.0 ed. Systat Software Inc., San Jose, CA Raguso RA, Willis MA (2003) Hawkmoth pollination in Arizona’s Thomson JD, Wilson P, Valenzuela M, Malzone M (2000) Pollen Sonoran desert: behavioral responses to floral traits. In: Boggs presentation and pollination syndromes, whit special reference to CL, Watt WB, Ehrlich R (eds) Butterflies: ecology and evolution Penstemon. Plant Species Biol 15:11–29 taking flight. The University Chicago Press, Chicago Torres C, Galetto L (1998) Patterns and implications of floral nectar Raguso RA, Light DM, Pickersky E (1996) Electroantennogram secretion, chemical composition, removal effects and standing responses of Hyles lineata (Sphingidae: Lepidoptera) to volatile crop in Mandevilla pentlandiana (Apocynaceae). Bot J Linn Soc compounds from Clarkia breweri (Onagraceae) and other moth- 127:207–233 pollinated flowers. J Chem Ecol 22:1735–1766 Torres C, Mimosa M, Galetto L (2013) Nectar ecology of Datura Raguso RA, Levin RA, Foose SE, Holmberg MW, McDade LA ferox (Solanaceae): an invasive weed with nocturnal flowers in (2003) Fragrance chemistry, nocturnal rhythms and pollination agro-ecosystems from central Argentina. Plant Syst Evol ‘‘syndromes’’ in Nicotiana. Phytochemistry 63:265–284 299:1433–1441 Rathcke BJ (1992) Nectar distributions, pollinator behavior, and plant Valtuena FJ, Ortega-Olivencia A, Rodriguez-Riano T (2007) Nectar reproductive success. In: Hunter MD, Ohgushi KT, Price PW production in Anagyris foetida (Fabaceae): two types of (eds) Effects of resource distribution on –plant interac- concentration in flowers with hanging droplet. Int J Plant Sci tions. Academic Press, San Diego, pp 113–138 168:627–638 Schemske DW, Bradshaw HD (1999) Pollinator preference and the Walter HE (2010) Floral biology of Echinopsis chiloensis ssp. evolution of floral traits in monkeyflowers (Mimulus). Proc Natl chiloensis (Cactaceae): evidence for a mixed pollination syn- Acad Sci 96:11910–11915 drome. Flora 205:757–763 Schlumpberger BO, Badano EI (2005) Diversity of floral visitors to Waser NM, Price MV (1981) Pollinator choice and stabilizing subsp. pasacana (Cactaceae). Haselto- selection for flower color in Delphinium nelsonii. Evolution nia 11:18–26 35:376–390 Schlumpberger BO, Raguso RA (2008) Geographic variation in floral Waser NM, Price MV (1990) Pollination efficiency and effectiveness scent of Echinopsis ancistrophora (Cactaceae); evidence for of bumble bees and hummingbirds visiting Delphinium nelsonii. constraints on hawkmoth attraction. Oikos 117:801–814 Collect Bot 19:9–20 Schlumpberger BO, Cocucci AA, More´ M, Se´rsic AN, Raguso RA Waser NM, Chittka L, Price MV, Williams NM, Ollerton J (1996) (2009) Extreme variation in floral characters and its conse- Generalization in pollination systems, and why it matters. quences for pollinator attraction among populations of an Ecology 77:1043–1060 Andean cactus. Ann Bot 103:1489–1500 Williams G, Adam P (2010) The flowering of Australia’s rainforests: Schmid S, Schmid VS, Zillikens A, Harter-Marquez B, Steiner J a plant and pollination miscellany. Csiro Publising, Australia, (2010) Bimodal pollination system of the bromeliad Aechmea p 105 nudicaulis involving hummingbirds and bees. Plant Biol Willmot AP, Bu´rquez A (1996) The pollination of Merremia palmeri 13:41–50 (Convolvulaceae): can hawkmoth be trusted? Am J Bot Siegel S, Castellan J (1988) Nonparametric statistics for the 83:1050–1056 behavioral sciences. McGraw-Hill, Singapore Wilmer P (2011) Chapter 8, Rewards 2: the biology of nectar in: Silva WR, Sazima M (1995) Hawkmoth pollination in Cereus pollination and floral ecology. Princeton University Press, peruvianus, a columnar cactus from southeastern . Flora United States 190:339–343 Wolff D, Braun M, Liede S (2003) Nocturnal versus diurnal Sipes SD, Tepedino VJ (2005) Pollen-host specificity and evolution- pollination success in Isertia laevis (Rubiaceae): a sphingoph- ary patterns of host switching in a clade of specialist bees ilous plant visited by hummingbirds. Plant Biol 5:71–78 (Apoidea: Diadasia). Biol J Linn Soc 86:487–505 Worley AC, Barrett SCH (2000) Evolution of floral display in Sletvold N, Trunschke J, Wimmergren C, Gren JA (2012) Separating Eichhornia paniculata (Pontederiaceae): direct and correlated selection by diurnal and nocturnal pollinators on floral display responses to selection on flower size and number. Evolution and spur length in Gymnadenia conopsea. Ecology 54:1533–1545 93:1880–1891 Yamamoto RT, Jenkins RY, Mcclusky RK (1969) Factors determin- Smith-Ramı´rez C, Ya´n˜ez K (2010) Digitalizacio´n de datos de ing the selection of plants for oviposition by the tobacco polinizadores de Chile, interaccio´n insecto-planta y distribucio´n hornworm Manduca sexta. Entomol Exp Appl 12:504–508 de insectos. Red Interamericana de Informacio´n sobre Yokota S, Yahara T (2012) Pollination biology of Lilium japonicum Biodiversidad var. abeanum and var. japonicum: evidence of adaptation to the StatSoft Inc. (2006) Electronic Statistics Textbook. Tulsa, OK. http:// different availability of diurnal and nocturnal pollinators. Plant www.statsoft.com/textbook/stathome.html. Accessed 25 Feb Species Biol 27:96–105 2013 Young HJ (2002) Diurnal and nocturnal pollination of Silene alba Stebbins GL (1970) Adaptive radiation of reproductive characteristics (Caryophyllaceae). Am J Bot 89:433–440 in angiosperms. I Pollination mechanisms. Annu Rev Ecol Syst 1:307–326

123